It is the purpose of this study to throw light on the degree to which the large corporations are linked among themselves through common control, community of interest groups, or more or less loose alliances.

It is of the very nature of the relationships which form the subject matter of this study that they are overwhelmingly qualitative in character. No statistical technique has been or is likely to be devised for reducing them to a quantitative scale. Furthermore informed observers will inevitably differ in their judgments about the weight to be attached to the various bits of evidence out of which a general picture must be pieced together. For this reason it is necessary to be as careful as possible in indicating the method of analysis which has been followed. Clearly no claim to unbiased accuracy can be set forth in a study of this sort; that fact alone puts the author under an obligation to present his material in a way to make critical appraisal possible and easy.

The kind of relationships which we are studying clearly have to do with the way in which corporations are managed and this in turn depends upon how and by whom they are controlled. How they are controlled may or may not be determined by their ownership. Consequently control is the central issue around which the study must turn.

2

Now it is a fairly simple task to classify corporations by the techniques employed in controlling them. The classification used by Berle and Means, while not exhaustive, is an excellent working scheme. They distinguish five major types, each one pretty much selfexplanatory: (1) control through almost complete ownership, (2) majority control, (3) control through a legal device without majority ownership, (4) minority control, and (5) management control. It is one thing, however, to be able to place a corporation in one or other of these categories and quite another to be able to identify and name the controlling individual or group. To a certain extent, to be sure, the two problems overlap. It is quite likely that if enough is known to place a corporation in one of the first four categories, enough will also be known to identify, at least in a general way, the controlling interest. This is not necessarily true, however, and in the case of the 5th category, it is likely not to be true. Since Berle and Means estimated that somewhere around one-half of the 200 largest nonfinancial corporations in 1929 were manage

1 Appendix 13, was prepared by Paul M. Sweezy.

Berle, A. A., Jr. and Means, G. C., The Modern Corporation and Private Property, 1933, ch. V.

Once the identity of controlling interests has been established it is possible to begin grouping companies together. This is, however, the most difficult task of all. Some corporations clearly belong together. For example, if one individual or well-defined group of individuals owns a majority of the voting securities of two or more concerns, then it will scarcely be denied that these companies should be placed together in what we may call a single interest group. We can safely say the same about any number of corporations which are completely under the control of the same interests, whatever the form of that control may be. But the concept of an interest group should surely comprise more than merely such corporations as are altogether under the same control. For example, if two brothers or close friends each own a business, and if at many points the policies of the two businesses are made in common, it would seem desirable to group the one with the other as belonging to the same interest group. Or if an investment banker promotes and takes a continuing and significant interest in several different concerns, it would appear that good grounds exist for putting these concerns into a single interest group. Most likely in the latter case the investment banker will be part of the management in each, sharing the control with others. We could generalize, then, and say that companies ought to be grouped together if, in the absence of counter-balancing factors, they have a significant element of control in common.

Does this mean that any two companies whose directorates interlock should be classed together in one interest group? The answer to this question is em

The history of every corporation has certain critical phases: organization and promotion, expansion, and possibly bankruptcy and reorganization. The role which certain individuals or groups play during these periods commonly determines their importance in more normal times. It is for this reason that it is so important to have a knowledge of historical facts.

4 In this connection, undoubtedly, the most valuable source of information is the magazine Fortune, which combines a high regard for accuracy with a special interest in personalities. On the other hand, there is very little to be found in the professional writings of economists and economic historians except in a few cases where the subject matter is specifically biographical.

6

phatically, "No." Anyone starting out on this principle would have little difficulty in putting all but a few of the 200 largest non-financial corporations into a single interest group. This fact is not without significance, but the classification achieved by this method would cover up the kind of grouping it is desired to disclose. For present purposes, material on interlocking directorates is unquestionably important, but it must be used with care and discrimination. Some general rules can be laid down, but in no case are they a complete substitute for knowledge of the relationships on which interlocking directorates are based. Interlocks may be classed as primary and secondary. A primary interlock exists between companies X and Y if a director of X, whose main business interest is with X, sits on the board of Y. If this same person also sits on the board of Z, then a primary interlock also exists between X and Z. These two relations, however, necessarily involve an interlock between Y and Z, and this we call a secondary interlock. It goes without saying that more weight should be given to primary than to secondary interlocks and that the latter should be interpreted only with caution.

More important in evaluating the significance of interlocking directorates is a knowledge of the general policies of the companies and individuals involved. Some firms and individuals regard the position of directorship as one of responsibility which involves their own reputations. They are not likely to assume such a responsibility unless they are in harmony with the general policy of the management of the company concerned and in a position to make their influence felt. This is clearly the case with the firm of J. P. Morgan & Co., for example. As a rule, a Morgan partner sits on the boards of only two or three large companies, frequently in related lines of activity. He is supposed to keep himself thoroughly informed and to take an active part in the affairs of these companies. When one considers the tremendous prestige which attaches to the Morgan name, it is easy to understand that the directorship of a Morgan partner is a fact of first importance in determining the orientation of a corporation. On the other hand, some individuals are perfectly willing to act as directors in a purely ornamental capacity, a function which in England is peculiarly reserved for members of the nobility. Directors, with no active business interests and no apparent asset except a name with prestige value,

See Appendix 12.

Cf. the statement made in a recent government investigation of railroads: "In investigations of control it has generally been the custom to lean rather heavily on interlocking directorates as a line of evidence. The present study prompts the view that such evidence can easily be overworked unless it is very exhaustively examined." Regulation of Stock Ownership in Railroads, 71st Cong., 3d sess., H. R. No. 2789, pt. 1, p LXXVI. This report will hereafter be referred to as Splawn Report: Railroads.

should always be regarded in this light unless there is specific evidence to the contrary.

It is obvious that multiple interlocks should be given more weight than single interlocks. In this connection, it is noteworthy that about half of the large companies in which J. P. Morgan & Co. is represented have two or more Morgan partners on their boards.

There are industrial and financial alliances which manifest themselves in other ways than through complete or partial common control. Most important are alliances based on banking and underwriting relations which do not result in formal interlocks. The connection between financier and manufacturer is generally not a casual one but a continuing one which gives rise to an active interest on the side of each party in the affairs of the other. Nevertheless, relations may remain entirely informal. For example, it was the general policy of Kuhn, Loeb & Co., under the leadership of Jacob Schiff, to eschew formal representation on the boards of its clients. Yet their responsibility for success was no less keenly felt. "Once a commitment had been made," Schiff's biographer comments, "the important task was to guide the borrower's financial projects in such a way as to promote their success. This essential service was not one which was legally due anyone concerned; yet it had to be rendered for the ultimate welfare of all. One way in which bankers can watch the interests of investors who look to them for guidance is to be represented in the management or board of directors of the concern for which they have issued loans. So far as Schiff was concerned he preferred, as a rule, that his firm should not be so represented. He felt that by personal conference and advice he could do as much as through formal representation."7 When relations are of the kind preferred by Schiff, they can only be recognized and evaluated by knowledge of the history of the companies involved.

Some alliances are of a kind which does not permit of generalization. Such, for example, is the close connection which has long existed between J. P. Morgan & Co. and the First National Bank of New York. It began as a personal relationship between the elder J. P. Morgan and the elder George F. Baker, but long since took on an institutional character. Outwardly this alliance manifests itself in close cooperation between representatives of the two concerns in the affairs of various third companies. Appointment to a partnership in J. P. Morgan & Co. is regarded as the most desirable form of promotion by junior officers of First National. Before the Banking Act of 1933 two Morgan partners were on the directorate of First National's

7 Adler, C. S., Jacob H. Schiff: His Life and Letters, 2 vols., 1928, vol. I, p. 27.

securities affiliate, the First Security Co., since dissolved. It would be misleading to call the MorganFirst National alliance unique, but it is certain that it would be difficult to fit into any general category. The list of such connections which defy generalization would be a long one; probably many exist which have altogether escaped the attention of the present writer. The best that can be done is to note them down and incorporate them in the general picture as they are discovered and checked.

From what has been said the reader will gather that the method followed in this study is thoroughly empirical and involves at every stage an exercise of practical judgment. An interest group is not a clear-cut concept which can be given concrete content according to mechanical rules. Accordingly the writer makes no claim to either completeness or finality. What follows should be regarded as tentative and subject to revision at many points if and when more adequate evidence is brought to bear on the problem. Only one general rule has been observed throughout and that is to disregard connections which are not based on pretty direct relations between two parties concerned.

There are, of course, no a priori limits to the scope which might be determined upon for this study. Ideally it should perhaps cover all significant interest groups judged by their relation to the economy as a whole. But such an ambitious project would take years to carry through, and the results would be difficult to present in a concise and readily intelligible form. Consequently, more or less arbitrary limits had to be imposed, firstly, on the segment of the economy considered; and, secondly, on the number of groups analyzed.

As to the first limit, the starting point was the list of the 200 largest nonfinancial corporations as of the end of 1935, presented and discussed in Appendix 10 above. The list had to be used before it had assumed final form so that there may be minor discrepancies between figures used in this section and those appearing in the final version of the list. The 200 largest nonfinancial corporations, for the purposes of this paper, then, include 107 industrials, 54 public utilities, and 39 railroads. It is inconvenient to handle the railroads as 43 separate companies since many of them are grouped together through minority stockholdings into large systems. In accordance with the procedure of the Splawn report on railroads, the bulk of the mileage has been grouped together into 13 major systems. This, of course, involves the inclusion of a number of smaller roads (not

8 Splawn Report: Railroads, part I, p. LII. The report names 14 major systems, but suggests (p. LI) that "the assignment of the Illinois Central to the Union Pacific system would perhaps be justified by reason of the fact that the latter owns by far the largest block of Illinois Central stock, representing 28.94 percent of the total." This assignment has been made here, and, consequently, the number of systems is reduced to 13.

in the Appendix 10 list) which belong to one or other of the major systems. The net result is that disucssion is limited to 13 major systems and 8 other roads with assets of $100 million or over.9

In addition to nonfinancial companies, it is necessary to consider at least banking companies in order to get a satisfactory view of the scope of important interest groups. This has been accomplished by including in the companies to be analyzed the 50 largest commercial banks as of the end of 1935.10

The total assets of the companies considered are set out in the following table:

It is possible to give a fairly accurate idea of the proportion of the total corporate assets of each class owned by the companies included in this table. According to figures presented in Appendix 11, at the end of 1933 the 104 largest corporations classified as "Manufacturing," "Mining and Quarrying," "Trade," and "Other," possessed 33.8 percent of the total corporate assets in these categories. The list is not quite the same as that for 1935, but the difference is of small order of magnitude. These classifications correspond to what have been summed up here under the heading "Industrials."

The 1933 figures indicate that the 96 largest corporations engaged in "Transportation and Other Public Utilities" owned 87.4 percent of all corporate assets in these fields. No precise breakdown between railroads and public utilities is available, but it is likely that the figure for rails should be somewhat higher and for utilities somewhat lower than 87.4 percent in their respective fields. In the case of rails, data compiled from the Splawn Report: Railroads show that the 13 major systems and 8 other roads included in the above table, owned at the end of 1929 about 95 percent of total railroad mileage. Assets figures would doubtless be roughly in proportion. Taking 95 percent as the correct figure for rails would mean, of course, that 75 percent would be about right for utilities.

In the case of banks it is possible to give a figure which is very nearly accurate. The 50 largest banks held, on December 31, 1936, deposits which amounted

There has been very little change in the composition of the major systems since the Splawn report. Nevertheless, in order to make the data as recent as possible, the grouping has been carried out in accordance with a chart compiled and published by Robert A. Burrows (Pittsburgh) entitled Inter-Relation and Capitalization of the Principal American Railroads-As of January 1, 1933. This chart is believed to be accurate and to embody all developments up to the time of its publication. In compiling asset figures for the systems, the assets of roads in which two systems have an equal interest have been divided between the two.

10 "Largest" by total resources as reported in Moody's Banks for 1936.

to 47.9 percent of the average deposits of all commercial banks for 1936.11 Assets figures would certainly not differ materially.

Summing up then, it may be estimated that the corporations included in this study own about 34 percent of the assets of all industrial corporations, 48 percent of the assets of all commercial banks, 75 percent of the assets of all public utilities, and 95 percent of the assets of all rails. It would probably not be denied that this sector of the economy is the seat of economic power out of proportion to its relative size.

The other limitation mentioned above, namely, the number of interest groups, has been more or less naturally dictated by the material itself.

From a careful company-by-company study there gradually emerged eight more or less clearly defined groups which so far overshadowed all the others that it seemed only logical to confine further attention to these eight.

It is manifestly impossible to rank these groups either by size or by influence. The interests of no two are equally divided among the different spheres of economic activity considered, nor are they at all strictly comparable from the point of view of the strength of the ties which bind them together. This point is important to emphasize. It if is kept in mind there is little danger of interpreting figures, despite their misleading appearance of precision, as more than general indicators of orders of magnitude.

The groups which will be considered may be designated for convenience as follows: (1) Morgan-First National, (2) Rockefeller, (3) Kuhn, Loeb, (4) Mellon, (5) Chicago, (6) DuPont, (7) Cleveland, and (8) Boston. The reasons for these particular labels should become clear in the course of the further discussion.

1. Morgan-First National.12-This group is for the most part based upon partial control by one or the other or more commonly both of the financial institutions after which the group is named. This partial control in turn is based upon long-standing financial relations and the very great prestige attaching to the Morgan and First National firms. Neither of these banking houses, however, operates through ownership to any significant extent. Some of the relationships which entitle corporations to membership in this group are more com

11 Data on the 50 largest are taken from the American Banker, January 19 1937, p. 11; and for all commercial banks from the Annual Report of the Federal Deposit Insurance Corporation for the Year Ending December 31, 1936, p. 125.

12 The banking act of 1933 enforced the divorce of deposit banking from underwriting. J. P. Morgan & Co. elected to continue in business as a deposit bank, and a new firm, Morgan Stanley & Co., Inc., was formed by a number of the partners of J. P. Morgan & Co, and Drexel & Co. (the Philadelphia branch of J. P. Morgan & Co.), to take over the investment banking business. Though J. P. Morgan & Co. and Morgan, Stanley & Co., Inc., are, of course, legally entirely separate entities, they have nevertheless been treated as one for purposes of this analysis.

14

The last two named are special cases. Glen Alden owns and operates the coal properties which once belonged to the Delaware, Lackawanna and Western Railroad. The ownership of the two are probably substantially identical, and we know that the D. L. & W. belongs to the extent of about 22 percent to the Bakers, the Vanderbilts, and the New York Central. Two representatives of the First National are directors of Glen Alden's subsidiary, Delaware, Lackawanna & Western Coal Co., which handles sales. St. Regis can be more advantageously discussed under utilities.

There is good reason to believe that all the companies which are listed as having Morgan-First National representation on their managements have more than merely formal relations with the two financial institutions. To review all the evidence would carry us much too far afield into the sphere of economic history. The list errs if at all, in the writer's opinion, on the side of understatement. These 13 industrials have combined assets of 3,920 million dollars.

The utilitiesincluded in the group are as follows:
American Telephone & Telegraph Co.
International Telephone & Telegraph Co.
Consolidated Gas Co. of New York.16
United Corporation group:

Commonwealth & Southern Corporation.
United Gas Improvement Co.

Public Service Corporation of New Jersey.
Niagara Hudson Power Corporation.
Columbia Gas & Electric Corporation.

13 This refers, as throughout this study, to the end of 1935.

14 Moody's Industrials, 1935, p. 1276. The railroads were obliged under the antitrust laws to divest themselves of coal properties.

15 Splawn Report: Railroads, part I, pp. 134–5.

16 Now Consolidated Edison Co. of New York.

Electric Bond and Share Group:17

American Power & Light Co.
American Gas & Electric Co.
National Power & Light Co.

Electric Power & Light Corporation. American Telephone & Telegraph has three directors in common with First National, but its informal relations with J. P. Morgan & Co. are probably even more important.18 Two Morgan partners are on the directorate of International Telephone & Telegraph.

The next group of companies, with which Consolidated Gas may well be considered, heads up into a superholding company called the United Corporation. United was formed in 1929 by J. P. Morgan & Co. and Bonbright & Co., acting in closest harmony.19 Its avowed purpose was to foster "closer relations among the great public utility systems in the east." 20 The first set of directors of United comprised five partners of a leading New York law firm and soon after its formation, "these directors resigned to make way for Messrs. Whitney and Gates of J. P. Morgan & Co. and Messrs. Thorne and Loomis of Bonbright & Co., Inc." 21 There is not the slightest doubt that these two companies were in sole control of later operations. The steps subsequently taken and the interrelations among companies in the United Corporation group are much too complicated to detail. In spite of the fact that stockholders in Consolidated Gas are insignificant, nevertheless this company is very closely tied in with the rest of the group, particularly through the fact that one man, Floyd Carlisle, is chairman of the boards of Consolidated Gas, Niagara Hudson, and St. Regis Paper Co. This, plus substantial stockholdings, also explains the inclusion of St. Regis Paper in the group.21a

The inclusion of the Electric Bond & Share System rests on less secure foundations than in the case of the United Corporation System. Nevertheless it is believed that the supporting evidence is amply convincing. Electric Bond & Share Co. was originally formed by General Electric Co. as a subsidiary to take over securities acquired by the latter in exchange for generating machinery and equipment." Though Gen

17 American & Foreign Power has been omitted from this study because all of its properties are held abroad. 18 The development of these relations has been traced in detail by the Federal Communications Commission in its investigation of the Bell System. See Federal Communications Commission, Special Investigation Docket No. 1, "American Telephone and Telegraph Company-Corporate and financial history", 3 Vols., Reports No. 22, 23, and 24.

19 For the story of the formation and development of United Corporation see "High Finance in the 'Twenties: the United Corporation," Columbia Law Review, May 1936, June 1936.

20 Columbia Law Review, June 1936, p. 936.

21 Ibid., May 1936, p. 787.

1a For interrelations within the United Corporation and Electric Bond and Share Group, see Inter-relation and Capitalization of the Principal Public Utility, Holding, Operating and Investment Companies, as of January 1, 1936. Compiled and published by R. A. Bunons, Pittsburgh.

22 Relation of Holding Companies to Operating Companies in Power and Gas Affecting Control, 73rd Cong., 2d Sess., H. R. No. 827, part 3, pp. 437 and ff. This report will henceforth be referred to as Splawn Report: Utilities

eral Electric divested itself of legal control in 1925, there was no change in management and there is no reason to suppose that the two concerns do not continue to cooperate as before. General Electric, it will be recalled, is one of the industrial corporations closest to the Morgan and First National banking houses. Furthermore Electric Bond & Share has had in the past, and may still have, relatively small minority holdings in stocks of United Corporation, American Superpower Corp., Commonwealth & Southern, Public Service of New Jersey, and Niagara & Hudson.23 "From the point of view of legal control," according to Bonbright and Means, "these stock interests of the Electric Bond & Share Co. in the United Corporation System are probably negligible. They become significant, however, by virtue of the fact that Electric Bond & Share Co. has long been closely affiliated with the banking house of Bonbright & Co., Inc., and they point strongly to the conclusion that the policies of the Electric Bond & Share Co. and of the interests controlling the United Corporation will be harmonious rather than antagonistic." 25 Nothing has happened since this was written to change this judgment.

24

The 12 utility companies included in the MorganFirst National group have combined assets of 12,191 million dollars.26

The assignment of railroad systems to the MorganFirst National group has been done sparingly. Only five major systems and one other road are included in the list, though an excellent case could be made out for according similar treatment to two more major systems and at least two other smaller roads. Those included are as follows:

New York Central System.27
Alleghany System.28

Northern System.29

Atchison System.30 Southern System.31 Western Pacific.32

Morgan and/or First National representatives partake in the managements of all the major systems listed, except Alleghany, and of Western Pacific. Financial

23 Bonbright, J. C. and Means, G. C., The Holding Company, 1932, p. 133. 24 Up to 1935, Sidney A. Mitchell, president of Bonbright & Co., was a director of Electric Bond & Share and three of its major subsidiaries.

25 Bonbright and Means, loc. cit.

26 The assets of United Corporation, American Superpower, Electric Bond & Share, and American & Foreign Power are not included in this total.

27 Includes New York Central; Delaware, Lackawanna & Western; and a one-half interest in Rutland.

28 Includes Chesapeake & Ohio; Missouri Pacific; Erie; New York, Chicago & St. Louis; Pere Marquette; Chicago & Eastern Illinois; Wheeling & Lake Erie; and a one-half interest in Denver & Rio Grande Western.

20 Includes Great Northern; Northern Pacific; Chicago, Burlington & Quincy; Spokane, Portland & Seattle; and Gulf, Mobile & Northern.

30 Includes only Atchison, Topeka & Santa Fe.

31 Includes Southern; and a one-half interest in Chicago, Indianapolis & Louisville. 32 Includes Western Pacific; and a one-half interest in Denver & Rio Grande Western